CCPS (Capacitor Charging Power Supply) / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

CCPS (Capacitor Charging Power Supply)

first previous 3 4 5 6 next last

Move Thread

LAN_403

Finn Hammer

Sun Sept 02 2007, 08:03PM

Registered Member #205 Joined: Sat Feb 18 2006, 11:59AM
Location: SkÃ¸rping, Denmark
Posts: 741

All,

Daniel and I have a problem understanding the relationship btwn. frequency and power transfer in the CCPS:

In this paper:

and this:

Well, same authors!

The transfer of energy is calculated as:

P = Fres * 0,5 * C * (2V)^2

4 times the energy stored in the resonant cap, per cycle.

Can anyone explain why it is 4 times the energy stored in the cap, and not just one times the energy stored in that cap?

We know it is true, because it is consistent with Microsim simulations.

Cheers Finn Hammer

Finn Hammer

Wed Sept 05 2007, 07:38PM

Registered Member #205 Joined: Sat Feb 18 2006, 11:59AM
Location: SkÃ¸rping, Denmark
Posts: 741

Insight is slowly trickling in, and I now realise, that it was probably wrong to assume up to 2kW into the jacobs ladder we made the other weekend.
Point is, the supply delivers constant _current_ into a load, so if this load is a low resistance, which I think a Jacobs ladder arc is, then the current is delivered into a low potential, and therefore the delivered power is also low.

Anyone know the resistance of a jacob ladder arc?

This weekend, we would like to test the supply under power, and we would like to do so into a resistive load.
Since we are expecting something in the vicinity of 15-20kW a water resitor comes to mind. A big barrell with salt water and copper electrodes would allow power measurements to be done with temperature rise as the calibration.

But which resistance will mirror the condition of the supply topping up, into a capacitor with 8kV on it?

Able to deliver 4 amps into the load, at a maximum voltage of 8kV, my shot at it is, that the resistor must be not over R=E/I 8000/4=2k. A little lower, really, to keep the load from rising to a voltage over the reach of the supply.

Does this sound reasonable?

Cheers, Finn Hammer

Hazmatt_(The Underdog)

Wed Sept 05 2007, 07:45PM

Registered Member #135 Joined: Sat Feb 11 2006, 12:06AM
Location: Anywhere is fine
Posts: 1735

Take some time to do a short circuit and open circuit test on the transformer, then go to the 4hv wiki under electronics and do the transformer analysis. That will give you the extracted elements of the transformer so you can model your problem.

You can find the impedance of the arc from the reflected load on the transformer. Essentially its short circuit at this point so you need to know the equivalent series secondary resistance at this point, but the transformer calculations will help with that.

Steve Ward

Wed Sept 05 2007, 08:52PM

Registered Member #146 Joined: Sun Feb 12 2006, 04:21AM
Location: Austin Tx
Posts: 1055

Finn, i think your 2k ohm load is the correct value (or at least pretty close) for doing a high power measurement. Keep in mind that this is more stressful than charging a cap, since when you charge a cap the average output power is 1/2 the peak output power when topping off the cap.

As to the jacobs ladder, why not just measure the voltage across the arc? I think it was in the few kV range for me (maybe 2kV).

Steve Conner

Wed Sept 05 2007, 10:44PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

I always wondered about the resistance of a Jacobs ladder. The maximum power transfer theorem says that you should be able to stretch an arc until it's pulling as much power as your supply can give. When it hits the maximum power point, it becomes unstable and goes out. That's what I believe anyway.

Steve Ward

Thu Sept 06 2007, 03:38AM

Registered Member #146 Joined: Sun Feb 12 2006, 04:21AM
Location: Austin Tx
Posts: 1055

Steve,

That is what i have measured in the past, but the voltage required to sustain longer growing sparks seems to go up with the square of the arc length (this is purely "by eye" type of measurement). In anycase, the voltage drop across the arc climbs steeply just before it breaks. It may present a load thats not too far off from charging a cap.

Daniel Uhrenholt

Thu Sept 06 2007, 02:31PM

Registered Member #125 Joined: Fri Feb 10 2006, 01:52PM
Location: Aalborg, Denmark
Posts: 155

Hi Finn.

I have to admit that I think a Jacobsâ€™s ladder is the way to go with our measurements at the moment.
But letâ€™s see what weekend bringsâ€¦

After some hard days with a lot of overtime at my place of
work, I managed to calculate a bit more on how many Jules/sec we can deliver with our CCPS as it is now.

Last time we had a Jacobâ€™s ladder on the supply, we had 400Vdc on the bridge. And with the transformers in parallel, we has 136nF in the tank and a Fo about 92kHz.
That gives about 4000 J/sec, but if we get the supply on three phase and 560Vdc in the bridge, we should get up to 7800 J/secâ€¦ Thatâ€™s almost twice the energy!!!

We need more C or a higher Fo

I canâ€™t prove that my calculations are correct, because my differential probe and my other probes are at your place

Cheers, Daniel

Finn Hammer

Sat Sept 08 2007, 08:00PM

Registered Member #205 Joined: Sat Feb 18 2006, 11:59AM
Location: SkÃ¸rping, Denmark
Posts: 741

All,

Today, scoping across the IGBT`s suggested that we`d probably better make us a low inductance bus:

Not that it solved our problems, because the whole thing still rings like the liberty bell:
TEK TDS2014, Metratek MX9003 diff probe, stock probe on low side gate.

(Yellow trace is voltage across a low side IGBT, blue is gate.)

We read up on RCD snubbers, and placed 1 across the buss between the lytics and the first IGBT.
The snubber consists of a 150nF CD942 cap, an IXYS DSEI 60-06A fast diode and a 1.5ohms resistor.
This brought os to this situation:

We then placed 2 more snubbers on the bus:
(The one to the left is the midle one, the 3rd is off the picture to the left.)

This resulted in this ringing waveform:

These shots are at really low voltages, we forgot to record the voltage, but below 50V.

When we turn the voltage up higher, the overshoot begins to climb up to over 50%, which is worrying, because it will correspond to more than 900V at 560V on the bus.

What?s left to do? Transorbs? Can we calculate the power in these overshots, so that we can also get an idea of how many tranzorbs we need?

Cheers, Finn & Daniel

Finn Hammer

Sun Sept 09 2007, 09:11AM

Registered Member #205 Joined: Sat Feb 18 2006, 11:59AM
Location: SkÃ¸rping, Denmark
Posts: 741

All,

If you will all have me excused for double posting, but I think everybody interested in H-bridge design should see this.

As you all know, the gate waveform is regularly displayed with great pride, "the squarer the better" seems to be the rule of the day.

But what about the voltage across swiching devices? We have just seen what ringing it can lead to.

Here are the waveforms showing Current in green, Voltage in blue and gate waveform (yellow), with a 22ohm gate resistor

Nice and calm, and the gate driver is happy, gate resistor is hot (2W) even the main IGBT`s are swallower.

Here is a shot at 250nS to show the calm voltage frontedge at 150V bus,

So with the new gate resistors in situ, the low inductance buss and the snubbers really came to play in harmony.
Looks like we are ready to ramp the voltage up to max, now.
Not today, though. There is an open house arrangement at the regional electricity plant, won`t miss that!

Cheers, Finn & Daniel

Steve Conner

Sun Sept 09 2007, 10:20AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Hi Finn & Daniel,

As far as I know, you are right. The answer to gate drive is not always "As fast as you can". You have to experiment to find out what is right for your system. I have built some (much smaller) DC-DC converters, and I spent a day fiddling with gate resistors and resistor-diode networks until I found the values that made the MOSFETs run coolest.

I'm also playing with those white Semikron IGBTs, and I've also noticed some fairly serious ringing on the bus even at low DC bus voltages. Maybe they are really fast

first previous 3 4 5 6 next last

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob